CN106557813A - The black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system - Google Patents
The black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system Download PDFInfo
- Publication number
- CN106557813A CN106557813A CN201610935847.8A CN201610935847A CN106557813A CN 106557813 A CN106557813 A CN 106557813A CN 201610935847 A CN201610935847 A CN 201610935847A CN 106557813 A CN106557813 A CN 106557813A
- Authority
- CN
- China
- Prior art keywords
- black
- technical problem
- supplying
- gas turbine
- energy system
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/10—Interfaces, programming languages or software development kits, e.g. for simulating neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
Disclosed by the invention is the black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system, is comprised the following steps:Black starting-up scene first to the distributing-supplying-energy system based on gas turbine as self-starting power supply is analyzed and classification;Then the more serious scene of selecting technology problem carries out analysis of Influential Factors, and technical problem mainly includes that idle-loaded switching-on line over-voltage amplitude and sky fill transformer excitation flow amplitude;Finally using error-duration model neutral net build black-start scheme technical problem assessment models, choose wherein best performance as black-start scheme technology evaluation model.The method of the present invention, simultaneously consider a plurality of circuit operation Overvoltage Amplitude and excitation surge current amplitude, and with reference to the characteristics of distributing-supplying-energy system, increased transformer capacity and load parameter is input into as the feature of neutral net on the spot, can obtain with more engineering adaptability, closer to the technical problem assessment Knowledge Verification Model of actual black starting-up situation.
Description
Technical field
The present invention relates to black-start scheme technical problem evaluation areas, more particularly to gas turbine distributing-supplying-energy system
Black-start scheme technical problem appraisal procedure.
Background technology
After black starting-up refers to that whole electric network fault is stopped transport, outer net help is independent of, by having self-startup ability in system
Unit self-starting after, drive the unit of other non self startings, gradually recover the process powered of whole system.On a large scale
System shutdown can cause the heavy losses of national economy, have a strong impact on stablizing for society.Within the very first time after system shutdown
Launch black starting-up, recovery system is powered as early as possible can effectively reduce the negative effect brought by system crash.When power failure is shortened
Between, reduce power failure cost angle, after it there is large-scale blackout, to go out practicable black-start scheme very necessary for rapid development.
Conventional self-starting unit includes water turbine set, diesel engine unit and gas turbine, and they have excellent self-starting
Ability, disclosure satisfy that the technical property requirements of black starting-up.But water turbine set is limited by geographical position, the quantity in large- and-medium size cities
Less, diesel engine unit has that pollution weight, efficiency are low.And the distributed system of gas turbine is currently based in country's encouragement
To be widelyd popularize under the policy of comprehensive energy supplying system development.Used as black starting-up power supply, gas turbine starts rapid, regulation spirit
It is living, additionally, when the distributed system based on gas turbine is as overall participation black starting-up, can also easily adjust in start-up course
Load on the spot is saved, to improve the technical problem index such as overvoltage, excitation surge current.Distributed system based on gas turbine will be in future
More and more important effect will be played in power grid"black-start".
The second step that black-start scheme is formulated will determine and be activated unit and transmission path.It is activated unit and generally chooses big
Capacity fired power generating unit, typically has multiple candidate targets;Transmission path is related to the selection of transmission line of electricity and approach transformer station, and scope is very
Extensively.This causes optionally to start scheme number numerous.When self-starting unit is gas turbine distributed system, due to
Which can reduce black starting-up spent time by the way of subregion starts and interconnects again, and each subregion has been required for respective black starting-up side
Case, causes the quantity of overall black-start scheme to be multiplied.
Actual black starting-up process is affected by the multiple technical problems of power system, wherein more serious problem includes circuit
Switching overvoltage and transformer excitation flow.Either the early stage of black-start scheme is formulated, or spot dispatch personnel are according to big
The practical situation that area has a power failure chooses preferred plan from alternative, it is necessary to carry out switching overvoltage and excitation surge current amplitude
Assessment verification, exclude amplitude exceed electrical equipment can tolerance range scheme, it is ensured that the feasibility of scheme.At present, generally adopt
Amplitude assessment is carried out with the mode of electromagnetism Transient State Simulation Software modeling and simulating, the method is modeled manually by technical staff, and needed
Want certain simulation time.The black starting-up candidate scheme numerous for number, carries out width one by one according to the method for modeling and simulating
Value verification, it will expend substantial amounts of energy and time.
Scholar both domestic and external proposes to obtain reflecting between black starting-up technical problem and relevant parameter by the method for neutral net
Relation is penetrated, then the relevant parameter in black-start scheme is updated in the mapping relations of neutral net, so as to rapid evaluation skill
The reasonability of art problem.But the research currently for the problems referred to above is concentrated mainly in switching overvoltage problem, is not directed to encourage
Magnetic shoves problem, the research of excitation surge current during black starting-up is then focused primarily upon and is qualitatively analyzed, lack the quantitative of amplitude
Forecast assessment.Additionally, the assessment verification to technical problem during black starting-up is concentrated mainly on water turbine set as self-starting electricity
The situation in source, the research invention to gas turbine distributing-supplying-energy system are less.
The content of the invention
It is an object of the invention to overcome the deficiencies in the prior art, there is provided the black starting-up of gas turbine distributing-supplying-energy system
Solution technique problem appraisal procedure, can be effectively realized the switching overvoltage and transformator to a plurality of circuit during black starting-up
Assessment verification while excitation surge current.
Technical scheme provided by the present invention is:
The black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system, comprises the following steps:
S1, to the distributing-supplying-energy system based on gas turbine as self-starting power supply black starting-up scene carry out classification with
Analysis;
The more serious scene of S2, selecting technology problem carries out analysis of Influential Factors, the more serious scene of the technical problem
Exceed predetermined threshold value J, sky including idle-loaded switching-on line over-voltage amplitude transformer excitation flow amplitude is filled more than predetermined threshold value K;
Threshold value J, K is set as needed;
S3, using error-duration model neutral net build black-start scheme technical problem assessment models technical problem is commented
Estimate.
Step S1 is specially:
Gas turbine distributing-supplying-energy system as can self-starting power supply, have difference in occurrence of large-area power outage
Application scenarios.Under different application scenarios, the electrical characteristic in black starting-up path has larger difference, needs to be ground respectively
Study carefully.
It is attributed to two classes using the distributing-supplying-energy system based on gas turbine as the black starting-up scene of self-starting power supply:Closely
Distance starts scene and remote startup scene;
Closely starting in scene, the power supply being activated is with distributed system distance less than setting value D, the transformation of process
Device quantity is less than setting value M, and it is unobvious that idle-loaded switching-on line loop operation overvoltage phenomenon and sky fill transformer excitation flow phenomenon, institute
State idle-loaded switching-on line loop operation overvoltage phenomenon and sky fills transformer excitation flow phenomenon and substantially do not refer to idle-loaded switching-on circuit mistake
Voltage magnitude fills transformer excitation flow amplitude less than predetermined threshold value K less than predetermined threshold value J, sky;Threshold value J, K and setting
Value D, M is set as needed;
Start in scene remote, the power supply being activated, is passed through more than or equal to setting value D with distributed system distance
Number transformer be more than or equal to setting value M, there is more serious line loop operation overvoltage phenomenon and transformer excitation flow
Phenomenon, the more serious line loop operation overvoltage phenomenon and transformer excitation flow phenomenon refer to idle-loaded switching-on line over-voltage
Amplitude exceedes predetermined threshold value J, sky and fills transformer excitation flow amplitude more than predetermined threshold value K.
The remote path schematic diagram for starting scene is as shown in Figure 2.
Under long distance power transmission scene, exist carries out the behaviour of idle-loaded switching-on to 110kV transmission lines of electricity and 220kV transmission lines of electricity
Make, the higher switching overvoltage of amplitude can be produced, when sky fills 110kV/220kV booster transformers, also amplitude can be produced higher
Excitation surge current.The present invention subsequently will be introduced for the remote scene that starts, but the method for the present invention is not limited at a distance
Start scene.
Step S2 is specially:
1) influence factor of idle-loaded switching-on line over-voltage is analyzed, affects the factor of nonloaded line closing overvoltage to include electricity
Source reactance, the reactance of transmission line of electricity and susceptance;
2) influence factor that sky fills transformer excitation flow is analyzed, affects the factor that sky fills transformer excitation flow to include electricity
Source reactance, transformer capacity;
3) analyze impact of the load to line over-voltage and transformer excitation flow on the spot.During black starting-up, from
Recover a small amount of load on the spot after starting unit starting and be able to maintain that its stable operation, a large amount of emulation find load restoration amount to operation
Overvoltage and excitation surge current amplitude have more significant impact.For distributing-supplying-energy system, with flexible energy supply mode,
The relatively general power system of its power load is more easy to control.Rational load restoration scheme is set during black starting-up, can
To improve related technical problem amplitude.Therefore, to the present invention research black starting-up problem, increase on the spot load value as one
Influence factor.
The factor of idle-loaded switching-on line over-voltage is affected mainly to include the ginseng such as source reactance, the reactance of transmission line of electricity and susceptance
Number, affects the factor that sky fills transformer excitation flow mainly to include the parameters such as source reactance, transformer capacity.
Understand with reference to above-mentioned analysis, the remote technical problem started in scene is mainly operated including 110kV transmission lines of electricity
Overvoltage, 220kV transmission lines of electricity switching overvoltage and 110kV/220kV transformer excitation flows three.Therefore, in the scene
In, the principal element related to above-mentioned technical problem includes source reactance, 110kV transmission line of electricity reactance, 110kV transmission lines of electricity electricity
Lead, transformer capacity, the reactance of 220kV transmission lines of electricity and 220kV transmission line of electricity conductance.
Additionally, during black starting-up, needing a small amount of load on the spot of recovery to maintain which stable after self-starting unit starting
Operation, a large amount of emulation find that load restoration amount has more significant impact to switching overvoltage and excitation surge current amplitude.For distribution
Formula energy supplying system, with more flexible energy supply mode, the power system that the electric load of output is relatively general is easily controlled
System.By arranging rational load restoration scheme during black starting-up, the technical problem amplitude of correlation can be substantially improved.Cause
This, to the present invention research black starting-up problem, increase on the spot load value as an influence factor.
Step S3 is specially:
1) choose suitable feature output and feature is input into:Feature is output as 110kV line over-voltage amplitudes, 220kV lines
Pass by voltage magnitude and 110kV/220kV booster transformer excitation surge current amplitudes;Feature input needs accurately reflect defeated to feature
The impact for going out, and easily obtaining, final feature input are chosen to be 7, are source reactance, load restoration amount, 110kV defeated respectively
Electric line length, 110kV transmission line of electricity models, transformer capacity, 220kV transmission line lengths, 220kV transmission line of electricity models;
2) training sample set and test sample collection are formed:Some exemplary values are assigned to each feature input variable and forms input
Sample space, sets up the phantom of gas turbine distributing-supplying-energy system black starting-up in PSCAD, in input sample space
Each sample emulate one by one, obtain the value of feature output, form overall sample set;To random after sample data normalized
It is divided into training and test sample collection;
3) training of neutral net and structure:The neutral net of different structure is trained by training set, is obtained not
With the feature input under structural network and the mapping relations of feature outlet chamber;
4) determine the neural network model of optimum structure, as black-start scheme technology evaluation model:By test sample
The performance of each network mapping relation of set pair is tested, and chooses error little, simple structure as optimum network.
Step S3 is described in further detail as follows:
A) choose suitable feature output and feature is input into
Feature output is the object of black-start scheme technology evaluation, i.e. 110kV line over-voltages amplitude, 220kV circuit mistakes
Voltage magnitude and 110kV/220kV booster transformer excitation surge current amplitudes.
The selection of feature input needs the factor for considering two aspects, and the input of one side feature is representative, can
The impact to exporting is accurately reflected, the concrete numerical value of another aspect feature input easily will be obtained.In step 2 to relative influence
Factor is analyzed, and wherein the reactance value of transmission line of electricity, susceptance value are respectively equal to taking advantage of for unit length reactance, susceptance and length
Product, and unit length reactance, susceptance are relevant with circuit model, therefore transmission line of electricity model and length are taken as feature input generation
For transmission line of electricity reactance and susceptance value, other influences factor is constant.
In sum, start scene for remote, the feature input of error-duration model neutral net is chosen to be 7, respectively
It is source reactance, load restoration amount, 110kV transmission line lengths, 110kV transmission line of electricity models, transformer capacity, 220kV defeated
Electric line length, 220kV transmission line of electricity models.
B) training sample set and test sample collection are formed
Rational span is arranged to each feature input variable first, several exemplary values are taken within the range,
The value of all variables is carried out into permutation and combination, input sample space is formed.
Then the phantom of gas turbine distributing-supplying-energy system black starting-up is set up in PSCAD, it is empty to input sample
Each interior sample is emulated one by one, records the value of three features outputs, and so as to obtain one, " seven features are defeated
Enter --- the output of three features " overall sample set.
Subsequently sample data is normalized, to reduce the unit and magnitude differences of sample set to training knot
The impact that fruit is caused, improves the training effectiveness of network.
Finally by normalization after overall sample be randomly divided into training sample set and test sample collection.
C) training and test of error-duration model neutral net
According to the training flow process of neutral net, it is trained by training sample set pair network, obtains being input into feature
With the neural network model of feature outlet chamber mapping relations, the flow chart of training method is as shown in Figure 3.In training process, training
Function is respectively adopted S type tangents frequently with the transmission function of Levenberg-Marquardt algorithms, network hidden layer and output layer
Function and S type logarithmic functions, train the Rule of judgment for terminating to be that error is less than set-point or iterationses exceed setting value.
After network training terminates, the input sample that test sample is concentrated is updated in network, obtains exporting test set, meter
The error between output test set and output sample set is calculated, the quality of network performance is judged according to error amount.
D) determine the optimum structure of neural network model
The structure of neutral net has a great impact to network performance.The neuron number and feature of input layer and output layer
It is input into the quantity exported with feature to be consistent, the hidden layer number of plies and hidden layer neuron number are typically determined by trial and error procedure.
Trial and error procedure needs training and the test for carrying out multiple network, by the comparison to error result, is meeting precision
Under the conditions of as far as possible select the neutral net of simple structure as optimal network, so as to avoid network from being absorbed in locally optimal solution, and subtract
The training time of few network.
The technical problem relevant parameter of black-start scheme is input in optimal neutral net model, you can directly obtain
110kV transmission line of electricity switching overvoltages, 220kV transmission lines of electricity switching overvoltage and 110kV/220kV transformer excitation flows
Amplitude, the highest tolerance range for whether exceeding electrical equipment according to amplitude can carry out the feasibility assessment of the black-start scheme.
Compared with prior art, beneficial effects of the present invention are:
1) present invention fully excavates its work in view of the fast-developing phenomenon of the distributing-supplying-energy system based on gas turbine
For the superiority of black starting-up power supply, for how having in the electrical network of multiple Distributed Integration energy supplying systems to black-start scheme technology
Problem is estimated verification, establishes the black-start scheme technology evaluation method based on neural network model.
2) in the selection being input into neural network model feature, the present invention considered the impact of each technical problem because
The characteristics of element and distributing-supplying-energy system, increased load restoration amount parameter on the spot so that the neutral net after training is more applicable
In the case of gas turbine distributing-supplying-energy system black starting-up.
3) present invention achieves the idle-loaded switching-on switching overvoltage amplitude and sky of a plurality of transmission line of electricity are filled static exciter and gushed
While stream amplitude, assessment verification, increased the transformer capacity parameter that can reflect excitation surge current amplitude, substantially increases black
Start the comprehensive and practicality of solution technique assessment.
Description of the drawings
Flow charts of the Fig. 1 for gas turbine distributing-supplying-energy system black-start scheme technical problem fast evaluation method.
Fig. 2 is the path schematic diagram for starting scene at a distance.
Flow charts of the Fig. 3 for the training method of neutral net.
Fig. 4 is output error change curve in neural network training process.
Fig. 5 is the difference schematic diagram between test sample desired output and network reality output.
Specific embodiment
With reference to instantiation, the present invention is described in further detail.
Such as Fig. 1, the black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system, including following step
Suddenly:
1) value of selected characteristic input
Assume the three-phase asynchronous switch-on during gas turbine distributing-supplying-energy system black starting-up.The numerical value choosing of feature input
Select situation as shown in table 1.The parameter of each aerial line is as shown in table 2.
1 feature of table is input into value table
2 aerial line parameter list of table
Model | Resistance Ω/km | Reactance Ω/km | Susceptance 10-6s/km |
LGJ-240/30 | 0.1181 | 0.399 | 2.86 |
LGJ-300/25 | 0.09433 | 0.412 | 2.76 |
LGJ-400/35 | 0.07389 | 0.404 | 2.82 |
2×LGJ-300/25 | 0.04717 | 0.311 | 3.62 |
2×LGJ-400/35 | 0.03695 | 0.308 | 3.67 |
2×LGJ-500/45 | 0.02956 | 0.304 | 3.70 |
2) test sample collection and training sample set are formed
According to the value in table 1 to different characteristic input, input sample space is formed, is built by PSCAD software emulations
Overall sample set, altogether comprising 2187 groups of samples.After normalized, 2000 groups of instructions as neutral net therein are randomly choosed
Practice sample set, remaining 187 groups used as test sample collection.
3) determine the optimum structure of error-duration model neutral net
Input layer is 7,7 feature inputs of correspondence, and output layer neuron is 3,3 feature outputs of correspondence.
The hidden layer number of plies is determined using trial and error procedure with neuron number.
The hidden layer number of plies is determined first.Find through the training and test of multiple network:When hidden layer is monolayer, error
It is larger, it is impossible to meet required precision;When hidden layer is three layers, training error meets required precision, and test error is excessive, occurs
Expired Drugs;When hidden layer is double-deck, training error and test error can meet required precision, and the training effect of network is most
It is good.Therefore the hidden layer number of plies is chosen for two-layer.The contrast of the error result that part is trained and tested is as shown in table 3, wherein, it is maximum
During error is both present in the training of 220kV line loop operation Overvoltage Amplitudes and tests.
3 different hidden layer number of plies error contrast tables of table
It is then determined that the number of hidden layer neuron.When two node in hidden layer of double implicit layer networks are close, net
Network training effect is best.Therefore, when neutral net is set up, the neuron number of every layer of hidden layer is equal for the present invention.Jing is sieved one by one
Choosing analysis, when the number of hidden nodes of network is 29, error is minimum, and training duration is also most short.Part relative analyses situation such as table 4
It is shown.
4 different hidden layer number of plies error contrast tables of table
In sum, final to determine that the hidden layer number of plies is two-layer, every layer of neuron number is 29.
4) interpretation of result
When neutral net hidden layer is bilayer, during per layer of 29 neuron, the change of the output error in network training process
Change shown in curve Fig. 4, output error is represented using mean square error.Do not occur locally optimal solution, output error one in training process
Directly continue monotone decreasing, training error has just reached 10-4 in 140 steps or so, and training performance is good.
Fig. 5 is the difference schematic diagram between test sample desired output and network reality output, using the shape of percentage error
Formula is represented.As a result show, excitation surge current, the amplitude of switching overvoltage and the PSCAD obtained by the neutral net for building is imitated
True result is basically identical.Transformer excitation flow concentrates on -0.5% with the percentage error of 110kV line over-voltage amplitudes and arrives
Between 0.5%, error very little;The percentage error of 220kV line over-voltage amplitudes is concentrated between -1% to 1%.It is maximum to miss
Difference absolute value is occurred in 220kV line over-voltage amplitudes, is 2.14%.
The percentage error of 220kV line over-voltage amplitudes is further analyzed, table 5 is 220kV line over-voltages
Amplitude error frequency distribution table, have more than 90% test sample error be located at it is interval (- 0.8%, 0.8%), be close to 75%
Test sample error be located at it is interval (- 0.5%, 0.5%);For training sample with test sample generally speaking, there is 96% mistake
Difference positioned at it is interval (- 0.5%, 0.5%), illustrate that neural metwork training is dry straight.
5 error frequency distribution table of table
Percentage error is interval | Training error frequency | Test error frequency | Global error frequency |
(- 1%, 1%) | 99.4% | 97.3% | 99.3% |
(- 0.8%, 0.8%) | 99.2% | 90.8% | 98.6% |
(- 0.5%, 0.5%) | 96.7% | 74.3% | 96% |
By the analysis to difference between desired output and reality output, the method by neutral net is illustrated, can be simultaneously
110kV transmission line of electricity switching overvoltages, 220kV transmission lines of electricity switching overvoltage and 110kV/ during black starting-up is predicted exactly
220kV transformer excitation flow amplitudes, so as to quickly be estimated to the feasibility of black-start scheme.
Claims (4)
1. the black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system, it is characterised in that including following
Step:
S1, the distributing-supplying-energy system based on gas turbine is classified and divided as the black starting-up scene of self-starting power supply
Analysis;
The more serious scene of S2, selecting technology problem carries out analysis of Influential Factors, and the more serious scene of the technical problem includes
Idle-loaded switching-on line over-voltage amplitude exceedes predetermined threshold value J, sky and fills transformer excitation flow amplitude more than predetermined threshold value K;
S3, using error-duration model neutral net build black-start scheme technical problem assessment models technical problem is estimated.
2. the black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system according to claim 1, its
It is characterised by, step S1 is specially:
It is attributed to two classes using the distributing-supplying-energy system based on gas turbine as the black starting-up scene of self-starting power supply:Closely
Start scene and remote startup scene;
Closely starting in scene, the power supply being activated is with distributed system distance less than setting value D, the transformator number of process
Less than setting value M, it is unobvious that idle-loaded switching-on line loop operation overvoltage phenomenon and sky fill transformer excitation flow phenomenon to amount, the sky
Carry closing line switching overvoltage phenomenon and sky fills transformer excitation flow phenomenon and substantially do not refer to idle-loaded switching-on line over-voltage
Amplitude fills transformer excitation flow amplitude less than predetermined threshold value K less than predetermined threshold value J, sky;
Start in scene remote, the power supply being activated is with distributed system distance more than or equal to setting value D, the change of process
Depressor quantity is more than or equal to setting value M, there is more serious line loop operation overvoltage phenomenon and transformer excitation flow is existing
As, the more serious line loop operation overvoltage phenomenon and transformer excitation flow phenomenon refer to idle-loaded switching-on line over-voltage width
Value exceedes predetermined threshold value J, sky and fills transformer excitation flow amplitude more than predetermined threshold value K.
3. the black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system according to claim 1, its
It is characterised by, step S2 is specially:
1) influence factor of idle-loaded switching-on line over-voltage is analyzed, affects the factor of nonloaded line closing overvoltage to include power supply electricity
The anti-, reactance of transmission line of electricity and susceptance;
2) influence factor that sky fills transformer excitation flow is analyzed, affects the factor that sky fills transformer excitation flow to include power supply electricity
Anti-, transformer capacity;
3) analyze impact of the load to line over-voltage and transformer excitation flow on the spot.
4. the black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system according to claim 1, its
It is characterised by, step S3 is specially:
1) choose suitable feature output and feature is input into:Feature is output as 110kV line over-voltage amplitudes, 220kV circuit mistakes
Voltage magnitude and 110kV/220kV booster transformer excitation surge current amplitudes;Feature input be chosen to be 7, be respectively source reactance,
Load restoration amount, 110kV transmission line lengths, 110kV transmission line of electricity models, transformer capacity, 220kV transmission line lengths,
220kV transmission line of electricity models;
2) training sample set and test sample collection are formed:Some exemplary values are assigned to each feature input variable and forms input sample
Space, sets up the phantom of gas turbine distributing-supplying-energy system black starting-up in PSCAD, to every in input sample space
Individual sample is emulated one by one, obtains the value of feature output, forms overall sample set;To being randomly divided into after sample data normalized
Training and test sample collection;
3) training of neutral net and structure:The neutral net of different structure is trained by training set, obtains different knots
The mapping relations of feature input and feature outlet chamber under network forming network;
4) determine the neural network model of optimum structure, as black-start scheme technology evaluation model:By test sample set pair
The performance of each network mapping relation is tested, and chooses error little, simple structure as optimum network.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610935847.8A CN106557813A (en) | 2016-10-25 | 2016-10-25 | The black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610935847.8A CN106557813A (en) | 2016-10-25 | 2016-10-25 | The black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106557813A true CN106557813A (en) | 2017-04-05 |
Family
ID=58443398
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610935847.8A Pending CN106557813A (en) | 2016-10-25 | 2016-10-25 | The black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106557813A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109917175A (en) * | 2019-03-11 | 2019-06-21 | 云南电网有限责任公司电力科学研究院 | It is a kind of for high anti-back-out when overvoltage method for quick predicting |
CN112290541A (en) * | 2020-10-15 | 2021-01-29 | 西安热工研究院有限公司 | Method for setting shunt reactor in black start process |
CN117728504A (en) * | 2024-02-18 | 2024-03-19 | 西安热工研究院有限公司 | Black start system and method for diesel-engine combined combustion engine |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103778230A (en) * | 2014-01-23 | 2014-05-07 | 华北电力大学(保定) | Online automatic generation method for black-start scheme |
CN104102954A (en) * | 2014-07-14 | 2014-10-15 | 南方电网科学研究院有限责任公司 | Distributive integrated energy supply system optimal configuration method considering black-start function |
CN104318317A (en) * | 2014-10-09 | 2015-01-28 | 华南理工大学 | Black-start scheme optimization method based on distributive integrated energy supply system |
-
2016
- 2016-10-25 CN CN201610935847.8A patent/CN106557813A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103778230A (en) * | 2014-01-23 | 2014-05-07 | 华北电力大学(保定) | Online automatic generation method for black-start scheme |
CN104102954A (en) * | 2014-07-14 | 2014-10-15 | 南方电网科学研究院有限责任公司 | Distributive integrated energy supply system optimal configuration method considering black-start function |
CN104318317A (en) * | 2014-10-09 | 2015-01-28 | 华南理工大学 | Black-start scheme optimization method based on distributive integrated energy supply system |
Non-Patent Citations (2)
Title |
---|
李膨源等: "基于反相传播神经网络的黑启动三相合闸统计过电压快速预测", 《华北电力大学学报(自然科学版)》 * |
陈国荣: "基于PSCAD的电力系统黑启动方案校验", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109917175A (en) * | 2019-03-11 | 2019-06-21 | 云南电网有限责任公司电力科学研究院 | It is a kind of for high anti-back-out when overvoltage method for quick predicting |
CN112290541A (en) * | 2020-10-15 | 2021-01-29 | 西安热工研究院有限公司 | Method for setting shunt reactor in black start process |
CN117728504A (en) * | 2024-02-18 | 2024-03-19 | 西安热工研究院有限公司 | Black start system and method for diesel-engine combined combustion engine |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107194055B (en) | Electric-gas interconnection system reliability modeling considering electric gas conversion device and evaluation method thereof | |
CN105337304B (en) | A kind of photovoltaic plant low voltage crossing data capture method | |
CN111555280B (en) | Elastic power distribution network post-disaster recovery control method based on electricity-gas comprehensive energy system | |
CN103944507B (en) | Photovoltaic-power-station low-voltage penetrating performance evaluation method based on inverter model test | |
CN109636009B (en) | Method and system for establishing neural network model for determining line loss of power grid | |
CN110417011A (en) | A kind of online dynamic secure estimation method based on mutual information Yu iteration random forest | |
CN106329516A (en) | Typical scene recognition based dynamic reconstruction method of power distribution network | |
CN103036230A (en) | Dynamic equivalence method of alternating-current-direct-current serial-parallel large power system based on engineering application | |
CN106992519B (en) | A kind of network load recovery robust Optimal methods based on information gap decision theory | |
CN105071771A (en) | Neural network-based distributed photovoltaic system fault diagnosis method | |
CN104318317A (en) | Black-start scheme optimization method based on distributive integrated energy supply system | |
CN103440497B (en) | A kind of GIS insulation defect shelf depreciation collection of illustrative plates mode identification method | |
CN106557813A (en) | The black-start scheme technical problem appraisal procedure of gas turbine distributing-supplying-energy system | |
Zhang et al. | Application of simulated annealing genetic algorithm-optimized back propagation (BP) neural network in fault diagnosis | |
CN104466959A (en) | Power system key line identification method and system | |
CN105676022A (en) | Long-line wind farm grid-connected resonance information extraction method | |
CN104319785B (en) | Source flow path electrical subdivision-based wind power system key node identification method | |
CN104934964A (en) | Power distribution network reconstruction and island division method containing distributed power supply | |
CN104463375A (en) | Power grid disaster recovery control model modeling method based on CIM standard | |
CN104393590B (en) | Electrical network Transient Instability pattern INTELLIGENT IDENTIFICATION method | |
CN103887792B (en) | A kind of low-voltage distribution network modeling method containing distributed power source | |
CN104102954B (en) | Distributive integrated energy supply system optimal configuration method considering black-start function | |
CN113139737A (en) | Comprehensive evaluation method for elasticity of electric power system of full-electric ship | |
CN105262108A (en) | Active power distribution network robustness reactive power optimization operation method | |
CN106526347A (en) | Digital-analog hybrid simulation-based photovoltaic inverter low voltage ride through evaluation method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170405 |
|
WD01 | Invention patent application deemed withdrawn after publication |